Unbounded heteroscedasticity in first-order autoregressive models and the Eicker-White asymptotic variance estimator
This paper focuses on first-order autoregressive models in which the noise variance increases without bound. Although this specification violates a standard assumption made in the relevant literature, namely that of bounded noise variance, it is proved that the well-known Eicker-White estimator remains a consistent estimator of the asymptotic variance of the OLS estimator.
Year of publication: |
2010
|
---|---|
Authors: | Kourogenis, Nikolaos ; Pittis, Nikitas |
Published in: |
Economics Letters. - Elsevier, ISSN 0165-1765. - Vol. 106.2010, 2, p. 84-86
|
Publisher: |
Elsevier |
Keywords: | Polynomial-like noise variance Asymptotic variance Order of variance growth Eicker-White variance estimator |
Saved in:
Saved in favorites
Similar items by person
-
Koundouri, Phoebe, (2015)
-
Statistical Modeling of Stock Returns: A Historical Survey with Some Methodological Reflections
Koundouri, Phoebe, (2012)
-
Can Statistical Models of Stock Returns "Explain" Empirical Regularities?
Koundouri, Phoebe, (2012)
- More ...